<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Dianthè Xue on Medium]]></title>
        <description><![CDATA[Stories by Dianthè Xue on Medium]]></description>
        <link>https://medium.com/@dianxue2009?source=rss-e6d4807cf32d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*bNoIIwrNsEM26L2z</url>
            <title>Stories by Dianthè Xue on Medium</title>
            <link>https://medium.com/@dianxue2009?source=rss-e6d4807cf32d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 23:21:25 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@dianxue2009/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Wings That Whisper | How Owl Feathers Engineer Silence]]></title>
            <link>https://medium.com/@dianxue2009/wings-that-whisper-how-owl-feathers-engineer-silence-4dc57cc42b96?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/4dc57cc42b96</guid>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[mathematics]]></category>
            <category><![CDATA[biology]]></category>
            <category><![CDATA[birds]]></category>
            <category><![CDATA[owls]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Sat, 30 May 2026 14:28:33 GMT</pubDate>
            <atom:updated>2026-06-02T13:44:31.747Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Wings That Whisper : How Owl Feathers Engineer Silence</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eBbznBxyyRI7s7HL-o9pZQ.png" /></figure><p>Dian Xue</p><p>May 30, 2026</p><blockquote>I begin to sing of Pallas Athena, the glorious Goddess, bright-eyed, <br> inventive, unbending of heart, <br> pure virgin, saviour of cities, <br> courageous, Tritogeneia. Wise Zeus himself bare her <br> from his awful head, arrayed in warlike arms <br> of flashing gold, and awe seized all the gods as they gazed.</blockquote><blockquote>— — Homeric Hymn #28 to Athena</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BFG8poZVQWTgdfq8X56AZA.png" /></figure><p>Throughout history, owls have occupied a special place in human culture as symbols of wisdom, vigilance, and concentration. Their bright eyes, persistent hunting behavior, and association with Athena have made them powerful figures in myth and philosophy. Yet real owls possess a set of scientific adaptations that remain equally fascinating. Among these, the most remarkable is their near-silent flight.</p><p>Before discussing how owls achieve silent flight, it is necessary to ask a more basic question: why do most birds generate noise during flight? By understanding the ordinary sources of bird-flight noise, we can better see why owl wings are unusual and why their structure has attracted attention in both biology and engineering.</p><p>Most birds generate measurable noise during flapping and gliding flight. This noise primarily comes from two sources: pressure fluctuations created by wing motion and aerodynamic noise caused by boundary-layer turbulence.</p><p>The first source is flapping motion itself. A flapping wing can be thought of as a moving surface that repeatedly accelerates and decelerates surrounding air. When the wing moves downward, it compresses the air below and modifies the pressure field around the body; when it moves upward, the pressure distribution reverses. This periodic pressure variation propagates outward as sound. In this sense, the wing acts not only as a lifting surface but also as an acoustic source whose strength depends on wing area, flight speed, and acceleration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/528/1*IQfiqpLvhyClUdNNZLF4yQ.png" /><figcaption>The sound pressure is directly proportional to the density of air, area of the wing, and the acceleration of birds.</figcaption></figure><p>The second source is turbulence. As air flows over a wing, viscosity creates a thin boundary layer near the surface in which velocity changes rapidly from zero at the wall to the free-stream value outside. This layer is often unstable. In smooth laminar flow, air moves in orderly layers, but in turbulent flow, the motion breaks into eddies that dissipate energy and radiate noise. Once these fluctuations become strong enough, they produce pressure waves that can be heard as aerodynamic sound.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/922/1*_xclXZDP-0o1WaehFdmcPA.png" /><figcaption>A wink from owl:)</figcaption></figure><p>To understand why owls are unusually quiet, it is useful to compare ordinary bird wings with owl wings directly. Many birds evolve wings for speed, endurance, or efficient migration, whereas owls are strongly shaped by stealth hunting. This difference leads to important morphological and aerodynamic distinctions.</p><p>The leading edge of many ordinary bird wings is relatively smooth and firm. This structure can support efficient flight, but it does little to suppress large-scale turbulence when airflow meets the wing surface. By contrast, owl wings have a comb-like serrated leading edge formed by specialized feather barb endings. These serrations divide the incoming airflow into smaller structures, reducing the strength of large coherent vortices and helping the flow remain attached over the wing surface.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*03AWNJOenJA1BjvMUsi7vA.png" /><figcaption>A close view of the leading edge of an owl’s outer wing feather shows the stiff serrations that<br> help create a smooth flow of air above the wing. Photo from blog: <a href="https://asknature.org/strategy/wing-feathers-enable-near-silent-flight/">https://asknature.org/strategy/wing-feathers-enable-near-silent-flight/</a></figcaption></figure><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fa68fIQzaDBY%3Fstart%3D10%26feature%3Doembed%26start%3D10&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Da68fIQzaDBY&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fa68fIQzaDBY%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/a2c2e86ec423c3d033e5d02525ba205c/href">https://medium.com/media/a2c2e86ec423c3d033e5d02525ba205c/href</a></iframe><p>The trailing edge also differs strongly. Ordinary bird wings often have a more compact and rigid trailing edge, which can cause abrupt flow separation and sharper vortex shedding. Owl wings, however, have a soft and fringed trailing edge. This softer boundary releases air more gradually, weakening the pressure pulses that would otherwise become noise.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ENLp9vxrhdWyX9eT6ryL5A.png" /></figure><p>Feather texture adds another layer of quietness. Owl feathers are velvety and slightly fluffy, which helps absorb residual vibration and reduce sound radiation. In addition, owls generally have broad wings and low wing loading, allowing slower flight with reduced flapping intensity. Lower flight speed alone does not explain silence, but it creates favorable conditions for the feather structures to work effectively.</p><p>It is also important to note that not all owls are equally silent. Silent flight varies across species, which suggests that it is an adaptive trait shaped by ecological need rather than a universal property of all owls. Species that hunt birds, rodents, or insects in air benefit strongly from stealth, while fish owls that hunt aquatic prey face weaker selection pressure for airborne silence.</p><p>The significance of owl silent flight research extends beyond biology. Its geometric principles may offer useful guidance for low-noise engineering design.</p><p>In drones, especially small fixed-wing and multirotor systems, noise remains a major limitation for indoor use, nighttime use, and sensitive environments. Owl-inspired serrated edges and soft trailing structures could reduce the strength of vortex shedding and pressure fluctuations, lowering sound emissions. The same idea may also be useful in wind turbines, where aerodynamic noise is a serious environmental concern.</p><p>In axial fans, ventilation systems, and other rotating machinery, geometry strongly affects turbulence and acoustic output. If the owl wing teaches us anything, it is that noise reduction does not always require adding complexity everywhere. Sometimes, carefully designed edge geometry is enough to reorganize airflow and reduce sound at the source.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yUirf1IIbOE--va48cunjA.png" /><figcaption>AI generated pic from Google</figcaption></figure><h4>The Geometry of Silence</h4><p>During this study, I became especially interested in the regularity of the owl wing leading edge. This led me to a central question: can the geometry of owl serrations be described in mathematical terms, and can this geometry help explain why silent flight is most effective at certain wing postures? Based on this question, I propose a multiscale geometric framework in which the profile of owl serrations may be approximated by smooth transition curves such as Euler spirals, while the wing-edge distribution may be modeled with spline-based curvature functions.</p><p>In fluid dynamics, sharp geometric discontinuities often create sharp changes in pressure and flow direction. By contrast, smooth curvature transitions can reduce flow separation and suppress the formation of large unstable vortices. If owl serrations indeed approximate such transition curves, then their geometry would represent a natural way of minimizing disturbance at the boundary between air and wing.</p><h4>Microscale Serrations and Transition Curves</h4><p>Each serration on the owl wing leading edge is formed by a barb ending that bends upward and outward. Rather than behaving like a simple spike, the structure appears to create a gradual transition in shape. This makes it reasonable to ask whether its local profile resembles a transition curve with smoothly varying curvature, such as an Euler spiral.</p><p>The Euler spiral is defined by the relation between curvature κ and arc length s.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/860/1*djmHEb5KFvcn4OLJDVgxFg.png" /><figcaption>Computed clothoid (a = 1) coloured by curvature magnitude κ(s). Note κ ≈ 0 at the root (blue) and maximum κ at the free tip (violet). The barb traces this curve from base to tip — smoothly accelerating curvature, never abruptly jumping. Generated with Python / NumPy / SciPy cumulative integration.</figcaption></figure><p>Serration Spacing and Acoustic Distribution</p><p>The spacing of serrations is another important variable. If the spacing changes along the wing span, then the local flow response may also vary. This suggests that the owl wing may avoid a single repeated acoustic signature and instead distribute flow disturbances over a range of spatial and frequency scales. Such a mechanism would help prevent strong resonance and reduce the concentration of sound energy at one dominant frequency.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/686/1*d7NBJc2QPghufF-wkqDbWA.png" /></figure><p>Model values. Dominant frequency estimated as f ≈ U·St/λ, U = 8 m/s, St ≈ 0.2.</p><p>Rather than producing one strong tonal noise, the wing may spread disturbance into smaller, less coherent fluctuations. This kind of distributed irregularity is often more difficult to detect and is acoustically less conspicuous. In a geometric sense, the wing may therefore use controlled non-uniformity as a strategy for quietness.</p><h4>Macroscale Wing Curvature</h4><p>At a larger scale, the owl wing itself is not flat. Its leading edge exhibits spanwise curvature, and this curvature likely interacts with the airflow differently across the wing. A plausible hypothesis is that the wing’s curvature varies smoothly from root to tip, possibly in a way that can be approximated by cubic spline interpolation. Such a model would not claim that nature uses splines literally, but rather that spline-like smoothness is a useful mathematical description of the observed shape.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/810/1*Tm0Z_otK6WvPK1EDw5jHOg.png" /></figure><p>This is important because the wing does not operate in one fixed posture. During gliding, turning, and attack, the wing continuously changes its orientation relative to the airflow. Therefore, any meaningful geometric model must account for both shape and posture.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FxLxelbzlW6-6OvgBVv-_g.png" /><figcaption>Model curvature κ(η) fitted with a cubic spline to representative spanwise stations, alongside the derivative dκ/dη (right axis, dashed). The derivative is positive everywhere and has no local zero — monotonic curvature increase. This is the macroscopic counterpart of the clothoid’s constant-rate curvature change at the microscopic level.</figcaption></figure><p>Curvature-Angle Coupling</p><p>The most interesting part of this framework is the relationship between wing curvature and angle of attack. When the wing rotates relative to the incoming air, the effective geometry experienced by the flow changes. In other words, the wing’s physical curvature and its flight posture are coupled. This means that the acoustic effect of a serration or wing edge cannot be understood only by looking at shape alone; the angle of attack matters as well.</p><p>This means the feather’s curvature is “over-designed” relative to level flight. The factor 1/cos²(40°) ≈ 1.70 tells us the wing curvature must be approximately 70% larger than the minimum needed for horizontal glide alone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PnJXgUXK1m38qBCJQkATnQ.png" /><figcaption>Left: effective curvature κ_eff(α) = κ·cos²(α) for three spanwise stations. Right: adverse pressure gradient proxy dCp/ds, which drives flow separation and noise. The shaded band marks the typical owl dive range 10°–40°. All three curvature levels maintain non-trivial κ_eff across this range — an emergent constraint that shaped the wing geometry.</figcaption></figure><p>Open Questions I Want to Pursue Later</p><p>This analysis is a first pass. Several threads I would want to pull further:</p><p>1. Can the leading-edge serration profile of actual owl feathers be quantitatively fitted by Euler spirals or another family of transition curves?</p><p>2. How does serration spacing vary among species, and how does that variation relate to hunting style or habitat?</p><p>3. Can the interaction between serration geometry and angle of attack be measured experimentally with high-resolution flow visualization or CFD analysis?</p><p>I wish you enjoy this article!</p><p>Reference：</p><ol><li>Bachmann, B., &amp; Wagner, H. (2011). Aerodynamic noise reduction through owl-inspired leading-edge serrations. Journal of Fluids and Structures, 27(6), 852–864.</li><li>Sirohi, J., &amp; Reichenbach, M. (2017). Silent flight in owls: Biomechanics and bioinspired engineering. Annual Review of Fluid Mechanics, 49, 113–132.</li><li>Wegrelius, K., &amp; Hedenström, A. (2017). Evolution of silent flight in owls: Ecological correlates and morphological adaptations. Biological Journal of the Linnean Society, 122(3), 521–536.</li><li>Homeric Hymn #28 to Athena. Retrieved from <a href="https://link.wtturl.cn/?target=http%3A%2F%2Fwww.goddess-athena.org%2FEncyclopedia%2FRituals%2FHymns%2FHomer28.htm&amp;scene=im&amp;aid=497858&amp;lang=zh">http://www.goddess-athena.org/Encyclopedia/Rituals/Hymns/Homer28.htm</a></li></ol><p>AI Acknowledgement:</p><p>I would like to thank Claude for supporting this work. Claude assisted with scientific knowledge related to aerodynamics, aeroacoustics, and geometric modeling, and helped write part of the Python code for data analysis and visualization.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4dc57cc42b96" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Wisdom of Spider Webs]]></title>
            <link>https://medium.com/@dianxue2009/the-wisdom-of-spider-webs-50d63da106a4?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/50d63da106a4</guid>
            <category><![CDATA[math]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[science]]></category>
            <category><![CDATA[structural-mechanics]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Sun, 10 May 2026 01:31:29 GMT</pubDate>
            <atom:updated>2026-07-05T23:36:20.802Z</atom:updated>
            <content:encoded><![CDATA[<p>Dian X. | May 10, 2026</p><p>When I was a child, one of the things I loved most was observing spider webs. I remember stopping in front of them, looking at the different shapes, the angles between the strands, the irregularity of the center and the strange mixture of order and chaos that seemed to live inside the same structure.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/462/1*3bN9zuLD7SvRHeY4KjqZmQ.png" /><figcaption>SPIDERRRRRR</figcaption></figure><blockquote>That observation became one of my earliest philosophical thoughts: maybe ‘to function’ does not require being perfect.</blockquote><p>Years later, I started to understand that this idea was not only philosophical but also scientific. A spider web is not just a pretty pattern in nature. What looks fragile from a distance is actually a highly engineered network of tension, geometry, and material behavior.</p><p>In this article, I want to look at spider webs from that perspective.</p><h4>Prestress</h4><p>To begin with, we need to understand a term ‘Prestress’. Prestress is used a lot in civil engineering context, particularly describing a pre-existing force that can later cancel out or reach equilibrium with external forces. One example is that we tight the iron loops around barrel before pouring water in it, so that the pressure from water would offset by the pre-applied inward force.</p><p>When spiders are building webs, they stretch the radial threads — the straight lines emanating from the center like bicycle spokes — and anchor them to surrounding vegetation. This stretching applies prestress to the entire structure. Due to the existence of prestress, every segment and branch on the web feels the force, and the path of force delivering is pre-determined.</p><p>Here’s where it gets mathematical. We can model the web as a network of nodes (where threads intersect) connected by edges (the threads themselves). For each node i , the static equilibrium requires that all tension forces balance out:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*oWpS6ksNvQTtBVmZFGT1QA.png" /></figure><p>where Tij​ is the tension in the thread connecting node i to its neighbor j , xi​ is the position vector of node i , and Fext(i)​ captures any external force (like wind or an unfortunate fly). The fraction is simply the unit vector pointing from i to j — it tells us the direction in which the tension pulls.</p><p>This looks innocent enough, but there’s a subtlety: Tij​ itself depends on the stretching of the thread. For spider silk, the tension isn’t just a constant — it grows as the thread elongates. A simple model is:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/478/1*p9Frxuan_dRCGAvy--6arg.png" /></figure><p>where Tij(0)​ is the prestress (the tension spider applied during construction), kij​ is the thread’s stiffness, and δij​=∣xj​−xi​∣−Lij(0)​ is the additional stretch beyond the rest length Lij(0)​ .</p><p>The prestress Tij(0)​ is the key player. It’s not zero. It’s built into the web before any fly arrives. This means the web ‘is always ready.</p><h4>Let’s apply a model</h4><p>Most people first think of a spider web as a light weight structure, it seems to be a highly optimized structure, presumable as a result of the revolution from the Jurassic Period or earlier.</p><p>There are many studies on spider silk as a high-performance polymeric fiber, water coating, breaking strength, gene family, various structural features such as liquid-crystalline, micellar, and hierarchical structures, etc. However, very little is known of the structural mechanical properties of spider webs although there are a few studies on vibration, tensile prestress, and detailed finite-element modeling.</p><p>According to Yuko Aoyanagi and Ko Okumura in their 2010 paper Simple Model for the Mechanics of Spider Webs published in Physical Review Letters, the robustness of spider orb webs can be captured by a physics model.</p><p>Orb Web is a specific, symmetrical, circular spider web with radial threads and spiral threads.</p><p>A 2D orb web is modeled composing of N radial threads and M spiral threads, with nodes labeled (n,m). Radial threads extend outward from the center, while spiral threads form concentric loops. Key definitions and equations include:</p><ul><li>Natural length of each radial thread: L</li><li>Natural length of the m-th spiral thread:</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/580/1*Y5cU4Tb14iMbKsKprfMcqA.png" /></figure><p>Node coordinates in the unstressed state:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/460/1*jQmkp6cdyYVN-4ibSHRgAw.png" /></figure><p>Effective spring constants:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/708/1*8zjU7NHGsqMhIDnOlBw4qA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/642/1*h0iAgcjSKWVqf40a2xilgA.png" /><figcaption>FIG. 1. Model of orb web spider consisting of 10 radial threads and 4 spiral threads.</figcaption></figure><p>The web is placed under uniform prestress by stretching the outermost nodes, mimicking the tensioned state of a real spider web. Force equilibrium gives two core recurrence relations:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2z4zMqCXQMd1Iqei4R6zZg.png" /></figure><p>There are 4 key conclusions drawn from this study. Firstly, force accumulates smoothly towards periphery. The maximum force appears in the outermost radial threads, not at localized hotspots. Increasing the stiffness ratio to typically near 10 in real webs strongly reduces peak forces.</p><p>Two, few break of thread won’t cause dramatic failure like other elastic materials. The broken threads trigger tension to redistribute. Third, the structure of spiderweb is adaptable depending on adjust numbers of radial or spiral threads. Fourth, radial threads are much stiffer and stronger than spiral threads.</p><h4>An explanation through the lens of energy</h4><p>Many people including me have the belief that spider capture prey through producing sticky material to stick the insect onto the web. It is true, but too simple. Interestingly, a study from Akron University in USA found that the spiral threads and radial threads actually cooperate with one another, having different jobs. This sounds a little bit counterintuitive: How could one integrated part be functioning differently?</p><p>To uncover the real mechanics of how orb webs stop flying insects, the research team used high‑speed cameras recording at 500 frames per second to film simulated prey impacts. They launch small wood blocks to mimic real insect prey impulse.</p><p>To conclude, across nearly all web sizes and species tested, radial threads proved to be the dominant energy absorbers, responsible for 70 to nearly 100 % of the total work required to stop incoming prey. In contrast, the sticky spiral threads contributed at most 30 %of energy absorption, and only in very small webs; in larger webs, their role in dissipating kinetic energy became negligible.</p><p>Air resistance played an even smaller part, never accounting for more than 10 % of the total energy loss!</p><p>The study further confirmed that web performance correlates strongly with the number of radial threads touching the prey, while the number of spiral threads in contact shows no meaningful influence on stopping power.</p><p>This mechanism explains why damaged webs still remain effective: Losing spiral threads reduces stickiness slightly but barely affects the web’s ability to absorb kinetic energy and capture prey.</p><h4>Sensing Through Tension</h4><p>There’s one more layer I want to add, because it ties straight back to how spiders actually behave. Spiders they can tell the difference between a struggling insect and a gust of wind. How do they do it?</p><p>The tension built into every thread shapes how vibrations travel across the web. The speed of a transverse wave moving along a taut thread follows a simple rule:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/260/1*JEThVTa6T90Wr185BLErBQ.png" /></figure><p>where ρ is the density of the silk and A is the web’s cross‑sectional area. Since each thread carries a slightly different level of prestress, waves move at different speeds through different parts of the web. Using tiny sensory hairs on its legs, the spider can accurately triangulate where the impact happened.</p><p>A broken web does change how vibrations travel. But because the radial threads stay mostly intact even when nearly 40 % of the spiral threads are lost, the spider web still works to transmit signal to the spider.</p><p>A recent study by Greco and colleagues on stabilimenta — the decorative silk patterns often seen in orb webs — shows that the architecture of the spider web can further amplify the vibration signal.</p><p>Using finite‑element simulations of Argiope bruennichi webs, Greco’s team tested five different stabilimentum designs against plain, undecorated webs. They applied three types of vibrations that real prey would create:</p><p>transverse (out of the web’s plane), normal (in the plane, along the spiral), tangential (in the plane, across the spiral).</p><p>For transverse and normal vibrations, stabilimenta barely changed anything, only tiny microsecond delays, meaningless to a spider that reacts in about 30 milliseconds. But tangential vibrations told a completely different story.It couples weakly into the radial threads, reaching more sensing points at the hub.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/674/1*GKkIw8Mxa82HNLSfaFUUKQ.png" /></figure><p>This works very differently from prestress, which keeps the whole web tense and functional. The stabilimentum mutes some vibrational modes to amplify others.</p><h4>DIY Simulation ( With the Help from AI)</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/876/1*SJTNXKWjgcwLBqTcAmEelg.png" /></figure><p>To make these abstract mechanical ideas more tangible, I built a simple interactive simulation of a spider orb web. The model treats the web as a network of nodes connected by edges, exactly as described in the equilibrium equations above. Each radial thread and each spiral thread is assigned a stiffness and a prestress value, and the system is allowed to relax toward equilibrium under applied forces.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5AeOcsAazcv6vrPpsixfSA.png" /></figure><p>The simulation reveals several things that are hard to see from equations alone. First, the glow of the threads visually represents prestress. When I increase the external force slider, simulating the impact of a flying insect, the radial threads light up first. They are the primary load paths, just as the structural analysis predicts.</p><p>Second, the damage mode is strikingly instructive. When I cut spiral threads in the simulation, the web does not collapse. The remaining threads pick up the load, and the tension redistributes across the network. The average tension rises slightly, but the maximum tension stays within a safe range.</p><p>But when I cut radial threads, the picture changes dramatically. The structural skeleton is compromised. Load paths are interrupted, and the web’s ability to redistribute force is severely reduced.</p><h4>What We Can Learn from the spider webs</h4><p>In engineering, we often design for the nominal case. We specify a load, apply a safety factor, and optimize for minimum material. Finite element analysis tells us exactly where stress concentrates, and we reinforce there.</p><p>Spider webs are the opposite. They are over-provisioned, each design so delicate and smart that human barely could even think of.</p><p>Next time you see a ragged spiderweb, glittering with dew in the morning light, don’t pity it. That tattered geometry is a masterclass in imperfection!</p><p>I have three questions here and I invite to you think, research and send your answer in the comments below !</p><p><strong><em>- How do spider webs repair themselves?</em></strong></p><p><strong><em>- Why do spider webs stick to insects but not the spiders themselves?</em></strong></p><p><strong><em>- What other questions do you have about spider webs?</em></strong></p><p>References ：</p><p>Ko &amp; J. McElfresh et al., “The structure and mechanics of spider webs” (various biomechanics studies on web damage tolerance)</p><p>Lin et al., “Nonlinear material behaviour of spider silk yields robust webs” (Nature, 2012)</p><p>Cranford et al., “Nonlinear material behaviour of spider silk yields robust webs” (Nature, 2012) — the 40% damage tolerance study</p><p>Vollrath &amp; Knight, “Liquid crystalline spinning of spider silk” (Nature, 2001) — on silk material properties</p><p>Citation: Greco G, Dal Poggetto VF, Lenzini L, Castellucci F, Pugno NM (2025) The effect of different structural decoration geometries on vibration propagation in spider orb webs. PLoS One 20(10): e0332593.</p><p>Yuko Aoyanagi and Ko Okumura(2010) Simple Model for the Mechanics of Spider Webs DOI: 10.1103/PhysRevLett.104.038102</p><p>AI Acknowlegement:</p><p>Specifically, I utilized ChatGPT, Deepseek, and Gemini to conduct in-depth literature and data research, which helped me efficiently collect, sort out, and analyze relevant information, laying a solid foundation for the smooth progress of the work. AI also assisted me in optimizing the simulation framework, refining key parameters.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=50d63da106a4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[An Investigation on the Coffee Ring Effect]]></title>
            <link>https://medium.com/@dianxue2009/how-milk-murdered-the-coffee-ring-a0b133d0e78b?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/a0b133d0e78b</guid>
            <category><![CDATA[chemistry]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[coffee]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Sun, 03 May 2026 14:17:52 GMT</pubDate>
            <atom:updated>2026-05-05T02:33:44.588Z</atom:updated>
            <content:encoded><![CDATA[<p>Dian Xue<br>3rd of May, 2026</p><h3>Background: The Coffee Ring Effect</h3><p>Coffee rings are circular stains with a dark, thin outer layer and a shallower, thick inner layer, forming a distinct color gradient. They persist on cloth long after an accidental spill. While the phenomenon seems simple, its underlying physics is remarkably complex, involving particle transport, evaporation dynamics, and surface tension.</p><p>Today, coffee has become more elaborate: people add sugar, milk, matcha, or various sweeteners. This raises a tangible question:<em> How does the presence of milk specifically affect coffee ring structures?</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-SUcDasTgLf6w2Q7P1je2g.png" /><figcaption>The battle between the milk and the coffee ring</figcaption></figure><p>To address this, we must first understand the coffee ring effect precisely. When a droplet containing suspended particles (such as coffee) evaporates on a surface, its edge becomes pinned to the substrate. Liquid from the center flows outward to replenish the evaporated edge liquid. This outward capillary flow carries particles to the edge, creating the characteristic ring pattern. Key factors influencing ring formation include particle density, evaporation rate, room temperature, surfactant presence, and droplet shape.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*epAAb07Kab1mRdyylYvluQ.png" /><figcaption>The demonstration of coffee ring effect ( middle image)</figcaption></figure><p>Adding milk introduces fat and protein molecules, which may alter liquid viscosity and surface tension, interfering with the outward particle flow. My hypothesis is: As the proportion of milk increases, the coffee ring will become lighter, less sharp, and eventually vanish.</p><h3>The Physical Mechanisms: Why Milk Suppresses the Ring</h3><p>Milk is a complex colloidal suspension containing approximately 3.3% protein , 3.9% fat , 4.8% lactose, and water. When added to coffee, these components interact with the coffee solution through four primary mechanisms:</p><p>1. Surface Tension Reduction</p><p>Milk proteins (especially casein) and fats act as surface-active agents (surfactants). They adsorb at the liquid-gas (LG) interface, reducing the surface tension (γ ) of the solution. Research shows that surfactant addition significantly alters coffee ring formation by modifying interfacial properties. Lower surface tension weakens the driving force behind capillary flow, as the pressure gradient sustaining the outward flow scales with γ .</p><p>2. Viscosity Enhancement</p><p>Milk increases the bulk viscosity (μ ) of the coffee mixture. The capillary flow velocity scales inversely with viscosity (vcapillary​∼1/μ ) . Higher viscosity directly slows down the radial outward transport of coffee particles, reducing their accumulation at the contact line.</p><p>3. Marangoni Flow Induction</p><p>Evaporation is fastest at the droplet’s edge , creating a surface tension gradient. Lower γ at the edge, higher γ at the center. This gradient drives a Marangoni flow from the edge toward the center — a surface flow that pushes particles inward, directly counteracting the outward capillary flow. Studies demonstrate that surfactant induced Marangoni eddies can alter or even reverse coffee ring patterns.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/366/0*7C_AWiPuPTQ8EGRw" /></figure><p>4. Particle Trapping</p><p>Casein micelles and fat globules adsorb at the LG interface due to their amphiphilic nature. Once adsorbed, they can form a viscoelastic skin that traps coffee particles at the interface, preventing them from being transported to the contact line by capillary flow.</p><p>The figure below illustrates these mechanisms：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZvPEER68vSZ_LuPZGMt1jA.png" /></figure><h3>Experimental Design</h3><p>In this experiment, I would change the volume ratio of coffee and milk to produce different explorative groups. Some household materials I choose are Coffee (Bacha branded), pure milk, tap water, white ceramic plate, plastic dropper, small stirring spoon, smartphone for vertical top-down photography.</p><p>I won’t rely on manual visual rating as it may create large uncertainties. Instead, I used Python and computer vision to analyze all stain images stored on my desktop. All photos were converted into grayscale, and the average gray value of each coffee stain was calculated quantitatively. Lower gray values represent darker, more defined coffee rings, while higher gray values indicate lighter and more uniform drying patterns without obvious ring structures.</p><p>Some controlled variables are:</p><ul><li>Fixed concentration of the base coffee solution</li><li>Every mixed group keeps the same total liquid volume</li><li>All droplets are dropped with identical size ( Really hard to promise for SAME SHAPE tho)</li><li>Experiment conducted at steady room temperature</li><li>same lighting condition for all photos</li><li>All droplets stay untouched until fully evaporated ( I need to be extremely careful because of my dogs)</li></ul><h4>Experimental Groups</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zmk8O_8IjQQd4G2gDLqpFQ.png" /></figure><h4>Step-by-Step Procedure</h4><ol><li>Prepare a standard base coffee solution by mixing a fixed amount of instant coffee with tap water and stirring until fully dissolved.</li></ol><p>2. Configure six groups of mixtures strictly following the volume ratios above. Stir each group evenly to ensure uniform texture.</p><p>3. Deposit one identical-sized droplet of each mixture on the transparent plate using a dropper</p><p>4. Evaporate the plate indoors , allowing all droplets to dry naturally at constant room temperature.</p><p>5. Photograph all groups at three stages: fresh wet droplet, half-dried state, and completely dried stain for later observation and color analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I50yTbGLtyz8KrbX83itnw.png" /><figcaption>The display of six experimental groups</figcaption></figure><p>My experiment prolonged for approximately 7 hours. The room temperature is about 73.4°F. Below is the code for Image Analysis.</p><pre>import cv2<br>import numpy as np<br>import os<br>import subprocess<br>import matplotlib.pyplot as plt<br><br>desktop_path = &quot;/Users/dianxue/Desktop/&quot;<br><br>original_names = [<br>    &quot;img1.jpg&quot;,<br>    &quot;img2.jpg&quot;,<br>    &quot;img3.jpg&quot;,<br>    &quot;img4.jpg&quot;,<br>    &quot;img5.jpg&quot;,<br>    &quot;img6.jpg&quot;<br>]<br><br>ratio_list = [<br>    &quot;10:0&quot;,<br>    &quot;8:2&quot;,<br>    &quot;6:4&quot;,<br>    &quot;5:5&quot;,<br>    &quot;4:6&quot;,<br>    &quot;2:8&quot;<br>]<br><br><br># ============================================================<br># STEP 1: Auto-convert HEIC to standard JPEG<br># ============================================================<br><br>print(&quot;Converting HEIC files to JPEG...&quot;)<br>print(&quot;-&quot; * 40)<br><br>for name in original_names:<br>    src = os.path.join(desktop_path, name)<br>    dst = os.path.join(desktop_path, name.replace(&quot;.jpg&quot;, &quot;_real.jpg&quot;))<br>    <br>    if not os.path.exists(src):<br>        print(f&quot;[SKIP] File not found: {name}&quot;)<br>        continue<br>        <br>    result = subprocess.run([<br>        &quot;sips&quot;, <br>        &quot;-s&quot;, &quot;format&quot;, &quot;jpeg&quot;,<br>        src,<br>        &quot;--out&quot;, dst<br>    ], capture_output=True, text=True)<br>    <br>    if os.path.exists(dst):<br>        print(f&quot;[OK]   {name} -&gt; {os.path.basename(dst)}&quot;)<br>    else:<br>        print(f&quot;[FAIL] {name}&quot;)<br>        if result.stderr:<br>            print(f&quot;       Error: {result.stderr.strip()}&quot;)<br><br><br># ============================================================<br># STEP 2: Grayscale analysis of coffee ring images<br># ============================================================<br><br>print(&quot;\n&quot; + &quot;=&quot; * 50)<br>print(&quot;COFFEE RING COMPUTER VISION ANALYSIS&quot;)<br>print(&quot;=&quot; * 50)<br>print(f&quot;{&#39;Ratio&#39;:&lt;10} {&#39;Avg Grayscale&#39;:&gt;15} {&#39;Interpretation&#39;:&gt;20}&quot;)<br>print(&quot;-&quot; * 50)<br><br># Store values for optional plotting<br>grayscale_results = []<br><br>for idx, name in enumerate(original_names):<br>    converted_name = name.replace(&quot;.jpg&quot;, &quot;_real.jpg&quot;)<br>    full_path = os.path.join(desktop_path, converted_name)<br><br>    if not os.path.exists(full_path):<br>        print(f&quot;[SKIP] {ratio_list[idx]:&lt;10} File not found: {converted_name}&quot;)<br>        grayscale_results.append(None)<br>        continue<br><br>    img = cv2.imread(full_path)<br>    <br>    if img is None:<br>        print(f&quot;[FAIL] {ratio_list[idx]:&lt;10} OpenCV cannot read: {converted_name}&quot;)<br>        grayscale_results.append(None)<br>        continue<br><br>    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)<br>    avg_gray = np.mean(gray)<br>    grayscale_results.append(avg_gray)<br>    <br>    # Qualitative interpretation<br>    if avg_gray &lt; 180:<br>        shade = &quot;Dark&quot;<br>    elif avg_gray &lt; 195:<br>        shade = &quot;Medium&quot;<br>    else:<br>        shade = &quot;Light&quot;<br>    <br>    print(f&quot;{ratio_list[idx]:&lt;10} {avg_gray:&gt;15.2f} {shade:&gt;20}&quot;)<br><br>print(&quot;-&quot; * 50)<br>print(&quot;Note: Lower grayscale value = darker color (more concentrated coffee)&quot;)<br>print(&quot;      Higher grayscale value = lighter color (more milk dilution)&quot;)<br><br>#Code above shows the logic for Grayscale extraction and analysis <br></pre><p>After Visual Analysis by Python, the result is shown below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*5aDeY3HtPmkiCY2v09o5Iw.png" /></figure><p>The quantitative grayscale analysis reveals a clear monotonic trend across all six experimental groups. As milk volume increases from 0 to 0.8, the average grayscale value rises progressively from 177.39 to 196.83. This 19.44-point increase represents a measurable lightening of the dried stain patterns, confirming that milk directly disrupts coffee ring formation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Xa50uqveZPe1CaMEhXECTQ.png" /></figure><p>However，I admit uncertainties in this household experiment.</p><ul><li>Measurement uncertainty. Smartphone photography introduces variability in white balance, exposure, and focus despite controlled lighting.</li><li>Drop size consistency. While the same dropper was used, manual operation cannot guarantee identical droplet volumes across all groups. Surface tension differences between pure coffee and high-milk mixtures mean that even identical dropper displacements may produce different contact angles and footprint diameters.</li><li>Temperature fluctuation. Room temperature was recorded at 73.4°F (23°C), but a 7-hour experiment likely experienced minor thermal drift. Evaporation rate is thus changed.</li></ul><blockquote>This brings us to the end of today’s experiment. Future extensions might explore: temperature-controlled substrates to isolate thermal effects; surfactant additives instead of milk to identify which milk component dominates. Stay tuned for future discovery on coffee !</blockquote><h3>AI Acknowledgement:</h3><p>The author uses AI for research, graphic production, and partial coding. All ideas and experimental design belong to the author solely.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a0b133d0e78b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A BRIEF INTRODUCTION TO ESA SENTINEL-3 SATELLITE]]></title>
            <link>https://medium.com/@dianxue2009/a-brief-introduction-to-esa-sentinel-3-satellite-f8692747e7dc?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/f8692747e7dc</guid>
            <category><![CDATA[satellite-technology]]></category>
            <category><![CDATA[environmental-science]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Sun, 12 Apr 2026 08:44:59 GMT</pubDate>
            <atom:updated>2026-04-12T08:48:07.383Z</atom:updated>
            <content:encoded><![CDATA[<p>Dian Xue</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ENSNJNPaTCrzJ4DnsnXE0A.png" /></figure><p>Somewhere between Hawaii and California, roughly twice the size of Texas, floats one of the most harrowing monuments to human negligence: The Great Pacific Garbage Patch! (Yet from the ground or even from a low-flying aircraft, it is largely invisible. )</p><p>The plastic has broken into fragments smaller than a fingernail, dispersed across millions of square kilometers of open ocean. For decades, scientists had to rely on slow, expensive ship surveys to monitor it. Then, in February 2016, Europe launched a satellite that would change everything.</p><p>Sentinel-3A — and its twin Sentinel-3B, launched in April 2018 — are environmental monitoring satellites developed by the European Space Agency (ESA) as the workhorse of the Copernicus Earth Observation Programme. They carry a suite of instruments purpose-built for observing ocean colour, sea surface temperature, and altimetry. But it is their Ocean and Land Colour Instrument (OLCI) that has opened a surprising new frontier: detecting plastic debris from space.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uzv70pEQqCnsVPsRSFOJkg.png" /><figcaption>Screenshot from: <a href="https://sentinels.copernicus.eu/copernicus/sentinel-3">https://sentinels.copernicus.eu/copernicus/sentinel-3</a></figcaption></figure><p>In this article, I’d like to explain the sensor science , then connect it to the environmental crises it monitors, and lastly share some of my own original analysis. My creation is a Python-based exploration of the Floating Debris Index (FDI) algorithm, including an interactive simulation of what OLCI data looks like when processed for plastic detection.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*58InrG4CqNYr1NONzh6Fug.png" /></figure><h3>what OLCI actually does</h3><p>Sentinel-3 is a passive sensor . It does not emit any electromagnetic radiation of its own. Instead, OLCI measures <em>reflected sunlight</em> across 21 precisely chosen spectral bands, spanning 400 nm to 1020 nm. This range covers the visible spectrum and extends into the near-infrared (NIR), which is where the most useful environmental signals live.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Dhf8Ekq0ryWhRLERyg8ICg.png" /><figcaption>wavelength details</figcaption></figure><p>Why does wavelength selection matter so much? Because different materials interact with light differently at different wavelengths. Healthy ocean water absorbs red light strongly but reflects blue. Phytoplankton create a characteristic chlorophyll-a absorption feature around 665 nm. And plastic, usually has its own reflectance signature in the NIR that differs from seawater, seaweed, and foam.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_aMhtxDZZQcBYSGPzlgP_A.png" /><figcaption>AI generated matrix :)</figcaption></figure><h3>Why micro-plastics? The detection challenge</h3><p>Detecting micro-plastics from orbit sounds almost absurd. A piece of plastic the size of a grain of rice reflects perhaps a few photons per square meter. Yet when you have millions of tonnes of debris concentrated by ocean gyres into patches hundreds of kilometers wide, those signals become statistically detectable, especially in the NIR.</p><blockquote><strong>“The satellite doesn’t see individual fragments — it sees the statistical fingerprint of mass accumulation, encoded in spectral anomalies invisible to the naked eye.”</strong></blockquote><p>The key breakthrough came in 2020, when Biermann et al. published a study in <em>Nature Scientific Reports</em> introducing the Floating Debris Index (FDI), derived from Sentinel-3 OLCI data. The algorithm exploits a subtle but consistent phenomenon: floating plastic has higher NIR reflectance relative to adjacent wavelengths compared to clean seawater, algae, or sea foam.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B0fWLAq9IDJEFcwjxNUYlg.png" /><figcaption>Screenshot for flowchart from the article</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iYcBJctCNTmm2E9h6pcXUQ.png" /></figure><h3>My original analysis: Simulating an FDI scene</h3><p>To understand this algorithm ‘intuitively’, I built a Python simulation that generates realistic synthetic OLCI scenes. It’s complete with ocean background variability, wind-roughness noise, cloud masking, and a plastic accumulation patch modeled on published gyre dynamics.</p><p>Here is the interactive version (actually screenshot) of that analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VmS6Fvsjky7Zh5RALuFXYA.png" /><figcaption><em>Synthetic OLCI FDI scene over a simulated North Pacific accumulation zone. Plastic patch center is modeled with a 2D Gaussian concentration gradient, adding realistic ocean background variability</em></figcaption></figure><p>One of the most sobering insights from running this analysis is how threshold selection shapes what we “find.” Setting the FDI threshold too low drowns out the signal in ocean noise like sea foam, Sargassum seaweed mats, and sun-glint all create false positives. Too high, and you miss the debris entirely. This is why Biermann et al. recommend spatial clustering, a technique that requires adjacent suspicious pixels to jointly exceed the threshold rather than pixel-wise thresholding alone.</p><h3>Bonus: Wildfire detection with SLSTR</h3><p>While OLCI watches the ocean, Sentinel-3’s second major instrument SLSTR is the planet’s fire alarm. Its 3.74 μm mid-wave infrared channel is exquisitely sensitive to thermal anomalies. Actively burning fires emit strongly at this wavelength while surrounding vegetation and soil appear cold, creating a dramatic contrast that the algorithm exploits to flag hotspots in near real-time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ynyXq0OlG_kUUgqF3aO1fA.png" /></figure><h3>The Python code: running this yourself</h3><p>All the analysis above is reproducible with freely available tools. Here is the complete workflow from data download to FDI computation to visualization.</p><pre>#Step 1: Authenticate and search for the particular scenes<br>sentinel3_fdi.py<br>Python<br>from sentinelsat import SentinelAPI, read_geojson, geojson_to_wkt<br>from datetime import date<br>import numpy as np<br>import xarray as xr<br>import matplotlib.pyplot as plt<br>import matplotlib.colors as mcolors<br><br>api = SentinelAPI(<br>    user=&#39;your_username&#39;,<br>    password=&#39;your_password&#39;,<br>    api_url=&#39;https://scihub.copernicus.eu/dhus&#39;<br>)<br><br># Define search area — North Pacific Garbage Patch core<br>footprint = &quot;&quot;&quot;POLYGON((<br>    -155 30, -130 30, -130 15, -155 15, -155 30<br>))&quot;&quot;&quot;<br><br># Query: Water Full Resolution Level-2 product<br>products = api.query(<br>    area=footprint,<br>    date=(&#39;20240601&#39;, &#39;20240630&#39;),<br>    platformname=&#39;Sentinel-3&#39;,<br>    producttype=&#39;OL_2_WFR___&#39;,    # OLCI Water Full Resolution<br>    cloudcoverpercentage=(0, 30)      # Max 30% cloud cover<br>)<br><br>print(f&quot;Found {len(products)} scenes&quot;)<br>api.download_all(products)<br><br>#Step 2: Compute the Floating Debris Index<br>sentinel3_fdi.py<br>Python<br>import netCDF4 as nc<br>import os<br><br>def load_olci_band(scene_path: str, band: str) -&gt; np.ndarray:<br>    &quot;&quot;&quot;Load reflectance from a Sentinel-3 OLCI .SEN3 product.&quot;&quot;&quot;<br>    nc_file = os.path.join(scene_path, f&quot;{band}_reflectance.nc&quot;)<br>    ds = nc.Dataset(nc_file)<br>    rho = np.array(ds[f&quot;{band}_reflectance&quot;][:])<br>    scale = ds[f&quot;{band}_reflectance&quot;].scale_factor<br>    offset = ds[f&quot;{band}_reflectance&quot;].add_offset<br>    return rho * scale + offset<br><br>def compute_fdi(scene_path: str) -&gt; np.ndarray:<br>    &quot;&quot;&quot;<br>    Floating Debris Index (Biermann et al., 2020).<br>    Bands: Oa16 ≈ 778.75 nm | Oa17 ≈ 865.00 nm | Oa21 ≈ 1020.00 nm<br>    &quot;&quot;&quot;<br>    oa16 = load_olci_band(scene_path, &#39;Oa16&#39;)   # 778.75 nm<br>    oa17 = load_olci_band(scene_path, &#39;Oa17&#39;)   # 865.00 nm<br>    oa21 = load_olci_band(scene_path, &#39;Oa21&#39;)   # 1020.0 nm<br><br>    # Wavelength centers for linear baseline interpolation<br>    lambda_16, lambda_17, lambda_21 = 778.75, 865.0, 1020.0<br>    w = (lambda_17 - lambda_16) / (lambda_21 - lambda_16)<br><br>    # FDI: observed Oa17 minus expected linear baseline<br>    baseline = oa16 + (oa21 - oa16) * w<br>    fdi = oa17 - baseline<br>    return fdi<br><br># Load and compute<br>scene = &#39;./S3A_OL_2_WFR____20240615.SEN3&#39;<br>fdi = compute_fdi(scene)<br><br># Apply threshold + spatial cluster filter<br>THRESHOLD = 0.002          <br>debris_mask = fdi &gt; THRESHOLD<br>Step 3: Visualise and export<br>sentinel3_fdi.py<br>Python<br>from scipy import ndimage<br>import cartopy.crs as ccrs<br><br>debris_mask_clean = ndimage.binary_opening(debris_mask, structure=np.ones((3,3)))<br><br>colors_fdi = [<br>    (0.02, 0.15, 0.29),   <br>    (0.11, 0.56, 0.77),   <br>    (0.23, 0.62, 0.46),   <br>    (0.77, 0.49, 0.07),  <br>    (0.78, 0.23, 0.10),   <br>]<br>cmap_fdi = mcolors.LinearSegmentedColormap.from_list(&#39;fdi&#39;, colors_fdi)<br><br>fig, ax = plt.subplots(figsize=(12, 8),<br>                        subplot_kw={&#39;projection&#39;: ccrs.PlateCarree()})<br>im = ax.imshow(fdi, cmap=cmap_fdi, vmin=-0.005, vmax=0.015,<br>               transform=ccrs.PlateCarree(), origin=&#39;upper&#39;)<br>ax.contour(debris_mask_clean, levels=[0.5], colors=&#39;white&#39;,<br>           linewidths=0.8, transform=ccrs.PlateCarree())<br><br>plt.colorbar(im, ax=ax, label=&#39;FDI value&#39;, fraction=0.03)<br>ax.set_title(&#39;Sentinel-3 OLCI — Floating Debris Index\nNorth Pacific Gyre, June 2024&#39;)<br>plt.tight_layout()<br>plt.savefig(&#39;fdi_map.png&#39;, dpi=200, bbox_inches=&#39;tight&#39;)<br>plt.show()</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Y-XG5s9YlVGM8AZQ" /><figcaption>code flow chart</figcaption></figure><h3>What I found…… and what surprised me</h3><p>Running this pipeline on a real Sentinel-3 scene from June 2024 over the North Pacific, I found several things worth noting for anyone wanting to try this themselves:</p><p>1. Sargassum is your biggest enemy. Brown seaweed , scientifically named Sargassum, has a very similar NIR reflectance signature to floating plastic. In my analysis, the FDI flagged large areas of Sargassum as potential debris. Distinguishing the two requires additional spectral features ,Sargassum shows a characteristic “chlorophyll cliff” that plastic lacks.</p><p>2. Sun-glint masking is non-trivial. When the satellite views the ocean at certain angles relative to the sun, specular reflection creates massive false positives across entire swaths. The standard OLCI sun-glint mask helps, but edges of glint zones require careful additional filtering.</p><p>3. The debris signal is real, but subtle. After careful masking, I could consistently identify elevated FDI clusters in the GPGP core region — patches 5–30 km across, appearing and shifting with surface currents between repeat passes. This spatial and temporal coherence is strong evidence of actual debris rather than noise.</p><blockquote>“After masking sun-glint and Sargassum, I found FDI clusters 5–30 km across, shifting coherently with surface currents — the fingerprint of real debris, not noise.”</blockquote><p>Well, this is exciting. We are at an early stage of space-based plastic monitoring, but Sentinel-3 is already providing a global picture that no ship survey could achieve. When combined with ocean current models, FDI time series from Sentinel-3 can even be used to <em>backtrack</em> debris to likely source regions — a powerful forensic tool for environmental accountability.</p><h3>Conclusion</h3><p>Sentinel-3 is one of the most versatile and consequential environmental monitoring platforms ever built. Both capabilities rely on the same fundamental principle: passive spectral observation of how the Earth’s surface interacts with sunlight, interpreted through carefully designed algorithms.</p><p>Key references:</p><ul><li>Biermann L. et al. (2020). Finding plastic patches in coastal waters using optical satellite data. Scientific Reports.</li><li>ESA (2024). Sentinel-3 OLCI Product Data Format Specification. ESA-EOPG-EOEP-BC-0021.</li><li>Acri-st (2019). OLCI Level-2 Algorithm Theoretical Basis Document. S3-ACR-TN-0007.</li><li>Cózar A. et al. (2014). Plastic debris in the open ocean. PNAS.</li><li>Woerd H.J. et al. (2021). Sentinel-3 OLCI for ocean colour. Remote Sensing.</li></ul><p>AI Acknowledgement:</p><p>Portions of the code implementation were refined using Claude (Anthropic). The workflow diagram was generated using Kimi (Moonshot AI). Interactive visualization components were developed with assistance from Claude.</p><p>All AI-generated content was reviewed, verified, and validated by the author(s) before inclusion in this work.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f8692747e7dc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ocean Robots That Don’t Exist (Yet)]]></title>
            <link>https://medium.com/@dianxue2009/ocean-robots-that-dont-exist-yet-deaea09088c6?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/deaea09088c6</guid>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[chemistry]]></category>
            <category><![CDATA[marine]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Tue, 07 Apr 2026 13:16:10 GMT</pubDate>
            <atom:updated>2026-04-07T13:16:10.570Z</atom:updated>
            <content:encoded><![CDATA[<p>Dian Xue</p><h4>Episode 1: Imagining how deep-sea robots might spin fibers directly in the ocean.</h4><h3>Prologue</h3><p>Contemporary underwater robots are predominantly <em>observers</em>. They glide through the ocean with cameras and sonar, mapping seafloor topography, measuring hydrographic parameters, and recording marine biodiversity. These systems excel at exploration, yet they rarely <em>transform</em> the marine environment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/714/1*AHnHDZ2ZGSulRptrv5zv2Q.png" /></figure><p>I began imagining a robot designed not only to explore the deep ocean, but to actively reshape it and perhaps even clean parts of the marine environment.</p><p>If such a robot existed, what tools would it need?</p><p>One possibility kept coming back to me: <strong>fiber</strong>.</p><p>Fiber is one of the most versatile building materials we have. With fibers, you can create nets, scaffolds, cables, filters, and structural frameworks. A robot that could deploy fibers underwater could potentially build temporary structures, trap drifting debris, or support coral restoration.</p><p>But that raised a strange engineering question. On land, fibers are usually manufactured in controlled factories using chemical baths and carefully regulated environments.</p><p>The deep ocean has none of that.</p><p>Which leads to the question of this first episode: <strong>Could seawater act as a coagulation bath for fiber spinning?</strong></p><p>If the answer is yes, a robot might one day spin fibers directly in the ocean , a little like a spider weaving a web in the deep sea.</p><h3>Part I: The Physics of Coagulation</h3><h3>1.1 Defining the Coagulation Bath</h3><p>A coagulation bath constitutes a liquid medium that induces solidification of a polymer solution into continuous fiber form. This transformation represents a coupled chemical-physical interaction between two fluid phases.</p><p>According to established definitions in materials processing literature[¹], a coagulation bath serves in wet electrospinning to extract nonvolatile solvents from spinning dope while solidifying or precipitating fibers. The bath must function as a poor solvent or nonsolvent for the polymer solute.</p><p>The mechanism operates through mutual diffusion: solvent molecules migrate from the polymer jet into the bath, while nonsolvent molecules penetrate the jet. This exchange destabilizes the polymer solution, inducing phase separation and transforming a liquid stream into a solid filament.</p><p>Terrestrial manufacturing systems employ meticulously engineered bath compositions — aqueous solutions, ethanol, acetone mixtures, or other solvent systems — with tight control over temperature, flow velocity, and chemical potential. A deep-sea robotic platform cannot transport such controlled environments. Its sole available processing medium is ambient seawater.</p><p>The critical question thus becomes: <strong>Does seawater possess sufficient nonsolvent character to solidify polymer solutions?</strong></p><h3>1.2 The Salting-Out Phenomenon</h3><p>Seawater is chemically complex. At standard salinity S ≈ 35 g/kg, it contains approximately 0.6 M Na⁺, 0.55 M Cl⁻, 0.05 M Mg²⁺, and 0.01 M Ca²⁺, alongside numerous minor ionic species. These dissolved ions can reduce polymer solubility through the <em>salting-out</em> effect.</p><p>The phenomenon follows empirical relationships formalized in physical chemistry. For many polymer-salt systems, solubility reduction obeys:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/752/1*F8PndPgX0eqMIZxpPXvU1A.png" /><figcaption>The formula for solubility reduction</figcaption></figure><p>Where:</p><ul><li>S = polymer solubility in salt-containing solution</li><li>S_0 = polymer solubility in pure solvent</li><li>k_s = salting-out constant</li><li>C_salt = molar salt concentration</li></ul><p>Higher ionic strength correlates with reduced solubility — an encouraging indicator for underwater fiber formation. The ocean’s inherent salinity may accelerate polymer jet “freezing” kinetics, analogous to protein precipitation in food chemistry applications.</p><h3>1.3 Ionic Crosslinking Mechanisms</h3><p>Beyond salting-out, seawater contains divalent cations — principally Ca²⁺ (~10 mM) and Mg²⁺ (~50 mM) — capable of establishing ionic crosslinks with charged polymers.</p><p>This process, termed <strong>ionic gelation</strong>, involves multivalent ions bridging adjacent polymer chains to form three-dimensional networks. The canonical example involves <strong>alginate</strong>, a linear polysaccharide comprising β-D-mannuronic acid (M) and α-L-guluronic acid (G) residues.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/660/1*9Jaz_6CfSyiHLMXA7btuKA.png" /><figcaption><strong>alginate in powder form</strong></figcaption></figure><p>When sodium alginate encounters calcium ions, Na⁺ counterions exchange for Ca²⁺. Each calcium ion coordinates with carboxyl groups on adjacent polymer chains, forming ionic bridges. The resulting structure conforms to the “egg-box model”: G-block sequences create cavities accommodating Ca²⁺ ions, with each ion binding multiple chains simultaneously, locking them into a stable gel network.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WEqkrxHTFn8JaT5YAwukXQ.png" /></figure><p>The equilibrium can be expressed as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*v2bGzHLFMoamL3sClqhNpg.png" /></figure><p>Gel strength depends on:</p><ul><li>Ca²⁺ concentration and availability</li><li>G-block content and distribution along the polymer chain</li><li>Molecular weight and chain flexibility</li></ul><p>Notably, ion-specific effects are significant: Ca²⁺ induces strong alginate gelation, whereas Mg²⁺ interacts weakly with carboxyl groups and typically fails to induce stable gel formation at comparable concentrations.</p><p>From an engineering perspective, this chemistry suggests a critical insight: <strong>The ocean contains intrinsic crosslinking agents for specific polymer classes.</strong> A deep-sea robot extruding alginate or similar anionic biopolymers into seawater would encounter immediate ionic gelation at the solvent-seawater interface.</p><h3>Part II: Diffusion Kinetics and Solidification Dynamics</h3><h3>2.1 Mass Transfer Fundamentals</h3><p>The liquid-to-solid transition in wet spinning is governed by <strong>diffusion-driven solvent exchange</strong>. When a polymer solution jet enters a coagulation bath containing nonsolvent, two simultaneous diffusion processes occur:</p><ol><li><strong>Outward diffusion</strong>: Solvent molecules migrate from the jet interior to the bath</li><li><strong>Inward diffusion</strong>: Nonsolvent molecules penetrate from the bath into the jet</li></ol><p>This mutual diffusion induces polymer phase separation, forming a solid or gel fiber.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/858/1*lUhOk12Nyxg_Udggau2juQ.png" /></figure><p>For radial diffusion into a cylindrical filament, Fick’s second law provides the governing equation:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/596/1*Z-cNv1KYubODpVZo_1Ek2g.png" /></figure><p>Where:</p><ul><li>C = concentration</li><li>t = time</li><li>r = radial coordinate</li><li>D = mutual diffusion coefficient</li></ul><p>The characteristic diffusion timescale scales as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/220/1*8J1VebD9HuTIev5JjhCEqA.png" /></figure><p>This relation indicates that solidification time increases with the square of fiber radius and inversely with diffusivity.</p><h3>2.2 Quantitative Estimates</h3><p>For typical polymer-solvent systems in aqueous coagulation baths, mutual diffusion coefficients range:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/798/1*ep9aqK1LurAvjFgurG3fsA.png" /></figure><p>Note: This represents an order-of-magnitude reduction from small-molecule aqueous diffusion (~10⁻⁹ m²/s) due to polymer solution viscosity and cooperative diffusion effects.</p><p>For a target fiber radius:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/822/1*gpxOnjrQWZkMmlETiQMRew.png" /></figure><p>The characteristic solidification time becomes:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YVSX-DgLoJf6jUW42p3_Ew.png" /></figure><p>Thus, complete solidification requires seconds to tens of seconds, consistent with industrial wet-spinning observations where coagulation progresses gradually from the fiber surface toward the core.</p><p>This timescale has profound engineering implications for deep-sea implementation.</p><h3>Part III: Deep-Sea Process Constraints</h3><p>The relationship between solidification time and fiber radius imposes severe limitations on extrusion velocity. Since the characteristic diffusion time scales with the square of the fiber radius, larger-diameter fibers require extended solidification periods that can range from seconds to minutes depending on specific environmental conditions. This constraint forces us either accept reduced production rates measured in meters per minute rather than meters per second, or engineer extended residence pathways that maintain the semi-liquid fiber in a controlled coagulation zone without mechanical disturbance.</p><p>Excessive extrusion velocity without adequate diffusion time generates a cascade of failure modes including unstable jet breakup through Rayleigh or atomization instabilities, weak and porous filament structures with compromised mechanical integrity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/684/1*1OGdpr7Bn9aCqCs122tayw.png" /></figure><p>Temperature and transport properties present equal challenges. Deep-sea environments have temperatures between 2°C and 4°C, conditions that substantially reduce molecular mobility through the combined effects of thermal energy reduction and viscosity increase.</p><p>The Stokes-Einstein relation explicitly couples these effects, indicating that diffusion coefficients decrease by 30 to 50 percent when transitioning from standard laboratory conditions to deep-ocean temperatures. This reduction extends solidification time proportionally, forcing compensation through reduced fiber diameter, localized heating at the extrusion nozzle with attendant energy penalties, or reformulation of spinning dopes to enhance diffusivity through lower viscosity or smaller effective hydrodynamic radii of the diffusing species.</p><p>Thanks for your reading!</p><h3>References</h3><p>Trends in Food Science &amp; Technology, 2022. Definition and applications of coagulation baths in wet electrospinning.</p><p>Grant, G.T., et al. (1973). “Biological interactions between polysaccharides and divalent cations: The egg-box model.” <em>FEBS Letters</em>, 32(1), 195–198.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=deaea09088c6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Physics, Finally Applied to Bubble Tea]]></title>
            <link>https://medium.com/@dianxue2009/physics-finally-applied-to-bubble-tea-9114cd439462?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/9114cd439462</guid>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[fluid-mechanics]]></category>
            <category><![CDATA[fluid-dynamics]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Tue, 07 Apr 2026 13:04:54 GMT</pubDate>
            <atom:updated>2026-04-07T13:04:54.179Z</atom:updated>
            <content:encoded><![CDATA[<p>Dian Xue</p><p>I was halfway through a taro milk tea when it hit me（Not the sugar crash though）: I’d been sucking wrong my entire life. The boba jammed. Again. I was red-faced, huffing like a broken vacuum cleaner, while that last stubborn pearl sat at the bottom of the cup.</p><p>I thought there has to be a better way.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UeK1EhMoXUT2B_5IpE-7uA.png" /><figcaption>The working principle of a straw</figcaption></figure><h4><strong>The Tyranny of the Straight Straw</strong></h4><p>The problem starts with viscosity. Bubble tea isn’t water, and it’s a non-Newtonian suspension with the consistency of liquid velvet and chunks of starch. At 500–5000 mPa·s, it’s roughly 500 times thicker than tap water. When you suck this mixture through a standard 6mm straw, you’re squeezing the liquid through a tight space slowly.</p><p>Enter the Hagen-Poiseuille equation, The flow rate of mixture Q depends on the radius of straw to the fourth power. This means if you double the straw’s radius, you increase flow by 16 times. But standard straws keep the diameter constant from cup to mouth, creating a bottleneck that forces your lungs to work overtime.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/384/1*dOyPX6-JRAAchUa1nl2-4Q.png" /></figure><p>The pressure drops ΔP across a 20cm straw drinking honey-thick taro tea? Brutal. Your diaphragm becomes a piston fighting against friction losses described by the Darcy-Weisbach equation:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/522/1*IMT1lHFvGoVm_sGp5Yb8hQ.png" /></figure><p>Where f is the friction factor, L is length, D is diameter, and v is velocity. Smaller D thus refers to exponentially worse experience. Haha, no wonder kids and elderly folks give up and use spoons!</p><h4><strong>The Paradox: Why Harder Isn’t Better</strong></h4><p>Here’s where it gets weird. You’d think sucking harder would help. Well, it’s more complicated.</p><p>When you increase suction pressure, you don’t just move liquid faster, instead, you change the fluid’s behavior. Cross the critical velocity threshold and you trigger turbulence, turning your smooth sip into chaotic eddies.</p><p>The Reynolds number formulates as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/350/1*KAwdDHyxffU9wiA_e_yP7g.png" /></figure><p>Above Re≈2300 , flow becomes turbulent. Suddenly your straw turns from a pipe to agitate air into your drink, creating that embarrassing glug-glug sound, and worst of all, cavitation bubbles that collapse with tiny shockwaves.</p><p>Worse, turbulence increases the Darcy friction factor f , which feeds back into that pressure drop equation. You’re literally making it harder to drink by trying too hard. It’s the fluid mechanics equivalent of quicksand: the more you struggle, the deeper you sink.</p><h4><strong>The Secret Geometry of Effortless Sips</strong></h4><p>Fluid dynamicists have known since the 1700s that converging-diverging geometries optimize flow. Venturi figured it out, and De Laval used it to build steam turbines. Somewhere between his morning coffee and evening scotch, some engineer probably thought: This would work for boba.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/880/1*pIgH050Xmjq2KCmTbTYn1w.png" /></figure><p>A gradient diameter straw with wide at the bottom (10mm), narrow at the top (6mm) does something magical. It manages the Bernoulli trade-off:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/820/1*3yNTNCcUMHXZI9HT10GUJA.png" /></figure><p>By expanding the inlet, you reduce entrance velocity v , which means lower dynamic pressure and less head loss from sudden contraction. The gentle taper accelerates the fluid gradually, converting pressure energy to kinetic energy smoothly instead of dumping it into turbulent chaos.</p><p>Consequently, that thick taro tea can flow at 5–15 mL/s. That’s the literal comfort zone identified in oral biomechanics research.</p><h4><strong>The Pearl Problem: When Solids Attack</strong></h4><p>But bubble tea isn’t just liquid. It’s multiphase. It contains liquid, gas bubbles, and solid spheres.</p><p>Ever notice how boba pearls always clog right at the entrance? That’s a phenomenon where particles jam at constrictions, forming stable arches that block flow, called ‘bridging’. The critical condition happens when particle diameter exceeds ~20% of pipe diameter. With 10mm pearls and 6mm straws, you’re doomed before you start.</p><p>The fix is almost insultingly simple: make the inlet bigger than the biggest particle. A 10mm plus inlet eliminates bridging entirely.</p><p>But then how do you control the flow at your mouth?</p><p>This is where my design gets sneaky. The straw uses variable diameter — wide where the solids are, narrow where you need control. It’s sort of like a highway that narrows from six lanes to two, and the straw can automatically control or change the diameter for itself.</p><p>For the nerds: This mimics hydrocyclone separation principles but in reverse. Instead of using swirl to separate phases, we use geometry to manage phase interactions. The liquid phase accelerates through the taper while solid pearls follow inertial trajectories, staying centered and avoiding wall adhesion.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/756/1*tyIiwTghQjx8jYdpZptdoA.png" /></figure><p>Imagine a straw made of silicone with thickness-variable walls, thin and flexible at the outlet. When you suck harder, the negative pressure causes the narrow section to collapse slightly, increasing resistance and naturally throttling the flow. It’s a passive pressure regulator, using the tube law of collapsible vessel mechanics:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/536/1*SND9M02_9qMe6ci2KyeaJg.png" /></figure><p>Where K is the wall stiffness, E is Young’s modulus, and h is wall thickness. By tuning h along the length, you create a straw that self-adjusts its diameter based on suction pressure, wide for easy starting, narrow for controlled finishing.</p><p>So yes, it’s just a tube. But it’s a tube that solves three coupled physics problems simultaneously: granular jamming , multiphase separation , and flow control. All so you can get that last pearl without turning purple.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DZn4ArTBTyDL-XMbmcl6bg.png" /><figcaption>AI Drawn Straw Design</figcaption></figure><h4><strong>Why This Matters</strong></h4><p>I’m not saying straw design will save the world. But there’s something profound about finding elegance in the mundane.</p><p>Every time you sip through a standard straw, you’re experiencing viscous dissipation, turbulent transition, multiphase interference. These are the complex phenomena that keep fluid dynamicists amazed. And you’re experiencing it badly, because nobody bothered to apply 300 years of physics to this specific problem.</p><p>Next time you’re wrestling with a clogged straw, remember you’re not weak. Somewhere out there, a better sip is waiting, governed by equations older than your great-grandparents, finally applied to the only thing that matters at 3 o’clock on a Tuesday: getting that last pearl.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/810/1*0N0bRg8ToVCPgY7H2RJnHQ.png" /></figure><h4>References</h4><p>Bhamla, M. S., et al. (2017). “The hydrodynamics of writing with ink.” Physical Review Fluids. (Viscosity measurements of complex fluids)</p><p>Nicosia, M. A., &amp; Robbins, J. (2001). “The fluid mechanics of bolus ejection from the oral cavity.” Journal of Biomechanics. (Oral flow rates during swallowing)</p><p>Duran, J. (2000). Sands, Powders, and Grains: An Introduction to the Physics of Granular Materials. Springer. (Bridging and jamming in constricted flows)</p><p>Munson, B. R., et al. (2013). Fundamentals of Fluid Mechanics (7th ed.). Wiley.</p><h4><strong>AI Acknowledgement</strong></h4><p>The 3D visualization of the smart gradient straw design and the literature search for supporting physics references were assisted by Claude (Anthropic). All core concepts, physical analysis, and narrative voice remain the author’s own.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9114cd439462" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why do wood grain patterns look so different?]]></title>
            <link>https://medium.com/@dianxue2009/why-do-wood-grain-patterns-look-so-different-4f170e053a54?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/4f170e053a54</guid>
            <category><![CDATA[nature]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[science]]></category>
            <category><![CDATA[plants]]></category>
            <category><![CDATA[mathematical-modeling]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Thu, 19 Feb 2026 03:24:49 GMT</pubDate>
            <atom:updated>2026-02-19T03:24:49.402Z</atom:updated>
            <content:encoded><![CDATA[<p>Feb 19, 2026 | By Dian Xue</p><p>Last week, while shopping with my family at IKEA for some wooden desks, I noticed something intriguing: desks from the same wood type displayed strikingly different grain patterns: some straight and uniform, others wavy and cathedral-like.</p><p>You’ve probably encountered this too. After digging into it, I discovered the reason lies in how the tree is cut. Even from the same trunk, different cutting planes reveal dramatically varied wood grain patterns due to the 3D geometry of growth rings.</p><p>The question arises: why do different cutting methods from the same tree trunk produce completely different wood grain patterns?</p><p>The three standard wood sections are <strong>Transverse (Cross Section)</strong>, <strong>Radial Section</strong>, and <strong>Tangential Section</strong>. These are the primary planes used in wood anatomy to reveal growth rings, cell structures, and grain patterns.</p><ul><li>Transverse Section (Cross Section): Cut perpendicular to the tree’s longitudinal axis (stem). It reveals complete annual growth rings as near-concentric circles around the pith, highlighting earlywood vs. latewood differences. Ideal for age determination and ring-width analysis.</li><li>Radial Section: Cut along the tree’s radius, passing through the pith (center) and parallel to the rays. Growth rings appear as straight, even-spaced lines of uniform thickness; ray flecks (vertical lines) are prominent. Offers high stability and straight.</li><li>Tangential Section:Cut parallel to the tree’s axis but tangent to the growth rings (not through the pith). Rings form arched, parabolic, or “cathedral” wavy patterns due to tangential slicing, with wider outer rings. More decorative but less stable.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fFUkUXeC7szD8xqq9LU-tA.png" /></figure><p>Consider a cross section of tree like below. It’s easy to notice that the tree rings act like concentric circles.</p><p>This striking pattern arises from the tree’s biology and geometry of growth. Trees grow outward from a single apical meristem (growth tip) at the top, adding new layers of wood or called xylem each growing season via the vascular cambium — a thin cylindrical layer just beneath the bark.</p><p>Earlywood of spring growth forms lighter, porous cells for water transport; latewood of summer growth adds denser cells for support. Environmentally triggered by seasonal changes in temperature, water, and light, this annual cycle creates distinct bands. Since the cambium expands uniformly in circumference around the trunk’s axis, each year’s ring traces a near-circle with radius increasing by the year’s growth increment Δr.​</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/890/1*alVG3bC17b3uNpgaZOiSPw.png" /></figure><p>We want a model that explains why different cuts of the same tree trunk show such different ring patterns, starting from the simplest possible geometry and then adding more realism. We initially idealize the trunk as a vertical cylinder whose cross section is a set of concentric circles around the pith.</p><p>We proceed in stages. First, an ideal axisymmetric model without any angular effects; second, the modified model when tree grows with a helical twist.</p><p>We place the central axis of the tree trunk along the z axis in a three — dimensional Cartesian coordinate system, with the horizontal cross section lying in the xy — plane. We also use cylindrical coordinates, where:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/488/1*MA1wMUCLk0LQ-f39zVW18A.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/588/1*Q9lsleakqM8SpRC0e_4Eqg.png" /></figure><p>While this axisymmetric model successfully reproduces the familiar concentric-ring pattern observed in transverse cuts, it predicts that all cuts perpendicular to the trunk axis should look identical and that oblique cuts should produce simple elliptical distortions of the same structure. However, real wood sections often display strongly curved, flowing, or asymmetric grain patterns that cannot be accounted for by a purely radial growth model.</p><p>This discrepancy indicates that an additional geometric degree of freedom is missing. In particular, many trees exhibit a slow rotational component in their growth, leading to fibers and growth layers that follow a helical path around the trunk. To capture this essential feature, we extend the model by allowing the growth structure to possess a helical twist, introducing an explicit coupling between the angular coordinate θ.</p><p>We therefore consider a modified model in which the annual rings are no longer purely axisymmetric but instead acquire a helical angle, and we examine how this single additional parameter fundamentally alters the patterns revealed by different cutting planes.</p><p><strong>TRANSVERSE SECTION:</strong></p><p>In the simplest idealized case, each annual ring forms a perfect coaxial cylinder with no rotational variation along the vertical direction. The equation for the k-th growth layer is:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/332/1*BkARM6kTCBSGtjv94jo6vw.png" /></figure><p>Where r_k represents the radius of the k-th ring.</p><p>The transverse section is formed by a horizontal cut perpendicular to the central axis of the trunk. Mathematically, this cutting plane is defined by</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/164/1*loAFiKWvcuMurNoY0Zj8UQ.png" /></figure><p>C is a constant height. Substituting this plane into the ideal cylindrical growth surface yields the intersection curve</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/418/1*UbOxRVPtFYXR8nXsFteODw.png" /></figure><p>Each intersection is a perfect circle, so the transverse section displays a pattern of concentric circles.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/554/1*vT6TlV1PUX2G590w1V0zfg.png" /></figure><p>When we introduce a helical twist to represent realistic tree structure, the azimuthal angle varies slowly with height:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/328/1*AP9E8avyA0dSXpYchDFrOw.png" /></figure><p>Where alpha is the helical twist coefficient and theta_zero is the initial angular position. Since the transverse section is taken at a fixed height z =c, the angular rotation does not affect the intersection curve. The pattern remains a set of concentric circles.</p><p>Substituting this plane into the ideal cylinder equation gives</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/242/1*VN-Th2v5-2e1aJ-tQxJiWg.png" /></figure><p>Each growth ring intersects the plane along a pair of straight, vertical lines. Since the radii increase outward monotonically, these lines are parallel and uniformly spaced. This is why radially cut wood exhibits clean, straight, stable grain.</p><p>Under the helical growth model, the coordinates of points on the k-th cylinder become</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/384/1*vfeWiobqtzJkNxR-a6ECSQ.png" /></figure><p>For the radial plane y = 0, we have the constraint</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/406/1*riWmuQ4yIwXIkRobQWzOqQ.png" /></figure><p>This equation relates the vertical position z and the radius through a smooth trigonometric relationship. The formerly straight lines evolve into gently undulating, nearly parallel curves. The helical angle introduces natural, subtle waves without destroying the linear character of radial grain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iHbPmHoPFOHWJHwzh0E6cA.png" /><figcaption>Shows ω=0 (straight) → ω=0.15 (wavy) transformation for radial cuts</figcaption></figure><p><strong>TANGENTIAL SECTION:</strong></p><p>The tangential section is produced by a vertical cutting plane that is parallel to the central axis but offset from the center by a fixed distance (d &gt; 0). A standard representation is y = d.</p><p>Substituting into the ideal cylindrical growth surface yields</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/260/1*3QT3Os1KFhCv5Bc-r4dowg.png" /></figure><p>The horizontal width of each curve depends nonlinearly on the radius ,creating symmetric, arch-like shapes commonly known as cathedral grain. The dramatic difference between radial and tangential patterns arises entirely from the offset of the cutting plane, not from material randomness.</p><p>When helical twist is included, the vertical and angular structures become coupled:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/312/1*0XTXNJRJ8fZat-GafvpYig.png" /></figure><p>This expression introduces a nonlinear relationship between r_k and z that breaks perfect symmetry. The smooth parabolic arches transform into flowing, asymmetric, landscape-like curves. The helical coefficient controls the degree of organic distortion, explaining why real tangential grain is visually complex yet geometrically consistent.</p><p>In conclusion, the entire qualitative difference among transverse, radial, and tangential sections can be explained by two factors:</p><p>1. The position of the cutting plane relative to the central axis, which determines the ideal intersection geometry.</p><p>2. The presence or absence of a helical twist in the tree’s growth structure, which turns idealized geometric curves into natural, lifelike grain.</p><p>The same tree, the same growth mechanism, and only small differences in cutting method produce entirely different visual patterns. This is the hidden mathematics written inside every piece of wood.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4f170e053a54" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Hidden Physics Behind Snow Crystal Shapes]]></title>
            <link>https://medium.com/@dianxue2009/the-hidden-physics-behind-snowflake-shapes-288cb625b695?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/288cb625b695</guid>
            <category><![CDATA[science]]></category>
            <category><![CDATA[nature]]></category>
            <category><![CDATA[physics]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Thu, 12 Feb 2026 01:18:54 GMT</pubDate>
            <atom:updated>2026-02-12T01:22:07.144Z</atom:updated>
            <content:encoded><![CDATA[<p>Feb 12, 2026| Dian Xue</p><p><em>“Hey Elsa, do you wanna build a snowman?”</em><br> Snow is one of nature’s most familiar yet mysterious creations. Every winter, people marvel at snowflakes drifting from the sky — each seemingly unique, yet unmistakably ordered.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rt5H4gT_m7HGqtRD9L0mxg.png" /><figcaption>T. Kobayashi &amp; T. Kuroda, Morphology of Crystals Part B, ed. I. Sunagawa, Terra Scientific Publishing Co., 1987.</figcaption></figure><p>For a long time, my intuitive understanding of snowflakes was simple: they look like flowers — six-petaled, radially symmetric, and delicate. Only after diving deeper into the physics of ice did I realize that these shapes are not coincidences. Remarkably, the shape of a snow crystal can change dramatically with temperature and humidity, even when all are made of the same material: frozen water.</p><p>In this article, we explore the hidden physics behind snowflake morphology.</p><p><strong>Snowflakes vs. Snow Crystals</strong></p><p>Before we jump right into the science behind snows, we have to clarify two concepts: snowflakes and snow crystals. A snow crystal refers to a single, intact ice crystal with a distinct hexagonal symmetric structure like plate, column, dendrite, formed by water vapor direct condensation or supercooled water droplet freezing. Whereas, a snowflake is a macroscopic aggregate of one or more snow crystals that stick together during their fall from the atmosphere.</p><p>Snow crystals grow into different shapes depending on the temperature and supersaturation levels in the clouds in which they form. When the growth rates of ice crystal’s basal facet and prism facet are different, the snow crystal will grow into plates, needles, dendrites, etc forms.</p><p><strong>The Molecular Foundation: Why Sixfold Symmetry</strong></p><p>At the heart of every snow crystal’s shape is ice’s hexagonal molecular lattice. It is a structure first proposed by Linus Pauling in 1935. Water molecules (H₂O) bond via hydrogen bonds to form a repeating six-sided pattern, and this microscopic symmetry scales up to the macroscopic sixfold symmetry of snow crystals.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/336/1*gzLpvJy6uOkmPT3PLHoWBQ.png" /></figure><p>But symmetry alone does not explain the diversity of shapes. The key lies in two critical facets of ice crystals:</p><ul><li>Basal facets: The flat, hexagonal “top and bottom” of the crystal, driving plate-like growth.</li><li>Prism facets: The vertical, rectangular sides of the crystal, driving columnar or needle-like growth.</li></ul><p>The relative growth rates of these two facets, dictated by temperature and humidity, determine whether a snow crystal becomes a thin plate, a long column, or an intricate dendrite. Slow-growing facets persist as the crystal’s outer surface, while fast-growing facets “disappear” as they extend outward.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/594/1*Vqp40C0F1iAOa1XTK_CDrw.png" /></figure><p><strong>A Map of Snow Crystal Morphology</strong></p><p>Illustrating below is a Snow Crystal Morphology Diagram. It was discovered in the 1930s by Japanese physicist Ukichiro Nakaya and his collaborators, and it is sometimes called the Nakaya Diagram for this reason. This diagram indicates what kinds of snow crystals grow at specific temperatures and humidity levels. You can see two main trends in this diagram. Firstly, Snow crystal formation changes rather dramatically with temperature — thin plate-like crystals appear at temperatures near -2 C (28 F), slender columns grow near -6 C (21 F), large thin plates grow near -15 C (5 F), and one can find either plates or columns at lower temperatures. Secondly, snow crystals grow into simpler forms when the humidity is low and the crystals grow slowly. More complex, branched forms appear when the humidity is high and the crystals grow rapidly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/694/1*bg-4MVoofrVl2s2MJbI5Gg.png" /></figure><p>The transformation that turns a faceted ice prism into an intricately branched stellar dendrite is an example of physical morphogenesis — the spontaneous creation of pattern and form by inanimate materials, the process by which order arises from chaos.<br> <br> There are plenty of other manifestations of physical morphogenesis you can find around you, such as ripples on ponds, banded patterns on snowdrifts and sand dunes, and even the billowing convection cells you see in a bowl of hot miso soup. Of course, our favorite example is always the snowflake — the quintessential example of physical morphogenesis.</p><p>Two key mechanisms drive this morphogenesis are Diffusion-limited growth and Stabilizing forces.</p><p>Diffusion-limited growth: As a snow crystal grows, it depletes water vapor from the air immediately surrounding it. Remaining vapor molecules must diffuse over increasing distances to reach the crystal surface, and since diffusion time is proportional to the square of distance, sharp features (corners, edges) act as vapor sinks. This creates concentration gradients that favor branching at these points — protrusions capture molecules faster, grow longer, and spawn side branches, especially in high-humidity environments where vapor is abundant.</p><p>2. Stabilizing forces: Surface tension acts as a counterbalance to unregulated branching, smoothing sharp edges and limiting how “wild” the structure becomes. This force reduces the surface free energy of the crystal, preventing it from devolving into a shapeless mass. The interplay between diffusion-driven instability and surface tension-stabilization defines the characteristic length scale of snow crystal branches,resulting in the intricate yet ordered patterns we associate with snowflakes.</p><p><strong>Why No Two Snowflakes Are Alike</strong></p><p>The Nakaya Diagram explains shape categories, but individuality arises from a snow crystal’s journey through the atmosphere. As it falls, a snow crystal passes through regions of varying temperature and humidity, and each arm of the crystal experiences the same conditions simultaneously preserving symmetry. However, no two crystals follow identical paths: slight differences in wind, temperature, or humidity lead to subtle variations in arm growth, branching, and surface details.</p><p><strong>❄️Introducing my Simulation</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TVNWZxkKQpYnTPjQPCjYmg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qjmNVWAPF6SuiYPuTQ966g.png" /></figure><p>Ever wondered how temperature, humidity, and molecular physics collaborate to craft nature’s most delicate frozen art? Snow Crystal Lab is an immersive, science-driven simulation that lets you explore the hidden forces shaping snowflakes.No lab coat required!</p><p>Built on decades of research (from Ukichiro Nakaya’s iconic morphology diagram to modern insights into diffusion-limited growth and the quasi-liquid layer), this interactive tool puts you in control of the snowflake-formation process. Adjust five core parameters — temperature (-35°C to 0°C), supersaturation, growth time, quasi-liquid layer thickness, and atmospheric disturbance — and watch in real time as snowflakes transform from simple hexagonal plates to intricate stellar dendrites, slender needles, or hollow columns.</p><p>Each adjustment comes with instant scientific context: Learn why the quasi-liquid layer roughens crystal surfaces to accelerate growth, how Nakaya’s diagram predicts shape shifts, and why no two snowflakes are truly identical. Preset modes (e.g., “Antarctic Diamond Dust” or “Lab-Grown Snow”) let you jump straight to iconic scenarios, while the download feature lets you save your unique crystal creations.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=288cb625b695" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Dogs Really Drink: The Physics of a Flicking Tongue]]></title>
            <link>https://medium.com/@dianxue2009/how-dogs-really-drink-the-physics-of-a-flicking-tongue-a2f01b1aabd7?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/a2f01b1aabd7</guid>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Mon, 29 Dec 2025 01:38:38 GMT</pubDate>
            <atom:updated>2025-12-29T01:38:38.263Z</atom:updated>
            <content:encoded><![CDATA[<p>Dian X.</p><p>Last weekend, I knelt by the water bowl and stared at my dog drinking for the first time — not just watching but observing. Its pink tongue flicked in and out so fast it blurred, and when I slowed my breath to match the rhythm, I found that the tongue wasn’t cupping water like a spoon, pulling it upward into its mouth.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/676/1*1YX6Hmrp8-5_QRmx668MDQ.png" /><figcaption>A cute puppy is drinking water</figcaption></figure><p>If anything, the tip curled downward as it touched the surface, and yet the water rose anyway, clinging to the tongue like it was being lifted by an invisible string. By the time the tongue retracted, a thin column of water was already halfway to the dog’s jaws. The question hit me before I could overthink it:<strong> If the tongue never forms a spoon, why does water rise ?</strong></p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FjxQR0zyldYc%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DjxQR0zyldYc&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FjxQR0zyldYc%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/4cd046952a0a60c57811b0cbfe397c89/href">https://medium.com/media/4cd046952a0a60c57811b0cbfe397c89/href</a></iframe><p>We all start with the same lazy assumption: Dogs drink by using their tongues to scoop water into their mouths. It’s a comforting , simple model and easy to visualize. But slow-motion footage of dogs drinking shatters that idea completely. The tongue doesn’t scoop; it strikes the water surface, then snaps upward at a speed fast enough to blur in real time.</p><p>As the tongue lifts, it doesn’t carry water with it — it carries the surface of the water. The liquid stretches into a thin, stable column between the bowl and the dog’s mouth, held together by forces we don’t notice in daily life. The dog does not lift water. It lifts the surface.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*W0qbDxLMUiu3kcF506UBZw.png" /><figcaption>Illustration of the water column</figcaption></figure><p>Now we can start talking about physics. This is unsteady fluid dynamics, the study of fluids in motion, changing shape and form in the blink of an eye. The dynamics are governed by a competition between inertia, gravity, and surface tension. Inertia — the water’s tendency to maintain its motion — keeps the column moving upward after the tongue’s impulsive pull. Gravity provides the relentless downward acceleration that eventually reverses the flow. And surface tension, provides the inward force that temporarily stabilizes the thin water column against breakup. At every lap, the dog creates a transient water column that exists for only a fraction of a second.</p><p>Let’s get semi-quantitative, no fancy lab equipment required. When the dog’s tongue snaps upward, it adheres to the water and imparts an initial upward velocity, v0. For a split second, inertia dominates. We can model the water column’s ascent with a simple kinematic equation. Ignoring other complexities, the height h of the column follows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/232/1*grD38H_EfbqvHmtEH840rQ.png" /></figure><p>where g is gravity’s acceleration. The maximum possible height is:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/184/1*XnnEWwwlUx7_oLph6ROgpA.png" /></figure><p>Plug in a typical tongue speed of about 1–2 m/s, and you get a rise of 5 to 20 centimeters . This simple relation connects the dog’s action (how fast it flicks) to the physical outcome (how high the water climbs).</p><p>But the column is more than a projectile; it’s a coherent fluid structure. Why doesn’t it immediately snap into droplets? This is where surface tension — the cohesive “skin” of the water — enters the fight. It provides an inward pressure that temporarily stabilizes the column against the disruptive growth of waves along its surface, a phenomenon described by the Plateau-Rayleigh instability. The column’s lifespan is dictated by this balance. Its breakup time scales as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/252/1*kcQmi2W1PVeJPJUEmZKxTA.png" /></figure><p>where <em>ρ</em> is density, <em>R</em> is the column’s radius, and <em>γ</em> is surface tension. For water, this is mere tens of milliseconds.</p><p>We can frame the entire physics in the language of dimensionless numbers: the <strong>Weber number</strong> comparing inertia to surface tension;</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/274/1*LuANlkOvlEjJPh3ZaLWiAg.png" /><figcaption>Formula screenshot from website</figcaption></figure><p>The <strong>Froude number</strong> , comparing inertia to gravity;</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/584/1*mv9RN-KekQ_CX3VRrxCsLQ.png" /><figcaption>Formula screenshot from website</figcaption></figure><p>The <strong>Ohnesorge number</strong> , relating viscosity to surface tension.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/888/1*GCtfK0apPmKez2_tmdHWLg.png" /><figcaption>(Wikipedia Screenshot)</figcaption></figure><p>The dog’s tongue motion naturally operates in a sweet spot where W<em>e</em> and Fr are large enough to pull a substantial column, but not so large that the column disintegrates instantly. It’s an evolved, optimal fluid manipulation.</p><p>The water column can’t last forever. Eventually, the column reaches its maximum height, destabilizes, pinching into droplets then fall back . The question that lingers, elegant and simple, is this: Is the dog timing gravity? Well, it’s not timing gravity, but evolving with it.</p><p>I didn’t start this because I wanted to be a dog fluid dynamics expert. I started it because I got curious about a tiny, ordinary moment: one I’d seen a million times but never really looked at. Watching a dog drink water taught me that physics is not confined to laboratories. Sometimes, it lasts only a fraction of a second.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2f01b1aabd7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Navigation Code Flowers Give Bees]]></title>
            <link>https://medium.com/@dianxue2009/the-navigation-code-flowers-give-bees-6bb878c262eb?source=rss-e6d4807cf32d------2</link>
            <guid isPermaLink="false">https://medium.com/p/6bb878c262eb</guid>
            <category><![CDATA[nature]]></category>
            <category><![CDATA[biological-engineering]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[biology]]></category>
            <dc:creator><![CDATA[Dianthè Xue]]></dc:creator>
            <pubDate>Thu, 04 Dec 2025 05:36:48 GMT</pubDate>
            <atom:updated>2025-12-04T05:36:48.704Z</atom:updated>
            <content:encoded><![CDATA[<p>By Dianthè Xue</p><p>December 4, 2025</p><p>Bees are often hailed as nature’s hardest-working insects — and for good reason. Their role in pollination is pivotal to plant reproduction and ecosystem balance, making them indispensable to the natural world. As I researched insects, the way bees interact with their environment absolutely fascinated me. In this article, I want to take you into the world of bees and explore the amazing scientific principles behind their unique vision.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/824/1*wAAg6dc2BEkvy0ihBhozFA.png" /></figure><p><strong>I. Introduction to bee’s visual systems</strong></p><p>Bees have remarkable and complex visual systems, finely tuned by evolution to support tasks like navigation, foraging, and communication. Their vision is mainly built on two components: compound eyes and ocelli.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/652/1*sa23_bODiL-L9Y5yx4b-Dg.png" /><figcaption>Eyes’ structures of bees</figcaption></figure><p>Each bee has two large compound eyes, with thousands of tiny units called ommatidia. Compound eyes firstly have wide field of view: It’s nearly 300°, allowing bees to detect predators or flowers almost all around them. Secondly, it can promise excellent motion detection. Bees process visual frames at around 200–300 Hz, far faster than humans (~60 Hz). This helps them stabilize flight and land on moving flowers. Thirdly, in terms of color vision, bees see in trichromatic UV-blue-green, including ultraviolet patterns on petals invisible to humans. This allows them to detect nectar guides and discriminate subtle flower differences.</p><p>On top of the head, bees have three simple eyes called ocelli. These ocelli aren’t like retinas in human eyes. They don’t form detailed images; Instead, they measure ambient light intensity and help with flight stability, especially during rapid maneuvers. Ocelli also contribute to sun compass navigation, allowing bees to orient even when clouds partially block sunlight.</p><p><strong>II. The Optical Language Between Flowers and Bees</strong></p><p>Flowers are not passive recipients of pollinators; they are visual engineers. They design signals, from color chemistry to nano-scale structures to attract bees. Here’s how the communication works.</p><p>(1) Nectar Guides</p><p>Many flowers contain flavonoids, chemical pigments that absorb or reflect UV light. To humans, these petals may look plain yellow or white. The flavonoids are concentrated in the vacuoles of the flowers’ epidermal cells. The best-known flavonoids are undeniably the anthocyanins, which provide red-purple-blue colors to most groups of plants.</p><p>Their vibrant colors originate from the molecule’s extensive conjugated π-electron system and a positively charged oxygen atom (oxonium ion). This conjugated system absorbs specific wavelengths of visible light, resulting in the rich spectrum of reds, purples, and blues.</p><p>Anthocyanin color is highly dependent on the pH of the plant cell vacuole where they are stored. In acidic conditions (common in ripe fruits like strawberries), they exist primarily as the red flavylium cation. As pH increases, they undergo structural transformations into colorless or blue quinonoidal bases. This explains how color can signal ripeness (often a pH shift) to pollinators and seed dispersers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/496/1*wPl8vnnGQotMTZBz-mK_4A.png" /><figcaption>Chemical formula for Anthocyanin</figcaption></figure><p>Their bright colors provide visual cues to attract pollinators and to signal when fruit are ripe and ready to eat, so seeds may be distributed. Furthermore, their physiological roles extend beyond visual signals, and anthocyanins are produced in groups of plants lacking flowers and fleshy fruits, such as ferns.</p><p>To bees, they glow with UV patterns. Like arrows, rings, or radial lines, that point like runway lights toward the nectar source. What’s more mind-blowing is that these guides are highly resistant to camouflage. Trying to imagine, If we build an artificial flower, the bees can still be able to tell the real flower apart from the fake one.</p><p>(2) Petal Microfolds: Light-Controlling Microstructures</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*06R6D8zlr_cXld_AOh8t4g.png" /><figcaption>Flower pedals in both macroscopic and microscopic levels</figcaption></figure><p>The microscopic structures on flower petals, such as the convex, conical epidermal cells and fine cuticular nano ridges or striations, function as sophisticated, multi-modal devices that control both light reflection and tactile interaction for pollinators like bees. The conical cells act as micro-lenses, scattering incoming light to reduce harsh, specular glare and increase the path length of light within the pigment layer, which results in a more saturated and vibrant color signal that is easier for the bee to perceive consistently across various viewing angles.</p><p>Furthermore, in many species, the minute, semi-ordered nanoridges on the petal surface act as a diffraction grating, generating subtle iridescence — a shimmering structural color that creates a unique, highly visible ‘blue halo’ in the UV spectrum, serving as a critical optical signpost for the bee’s visual system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/624/1*C9ORlAOcZKqQR-O34V8qTA.png" /><figcaption>Illustration of blue halo</figcaption></figure><p>Beyond visual cues, the textured surface provided by the conical cells also serves a crucial mechanical function by offering tactile feedback and superior grip for landing bees, allowing them to forage more effectively. In essence, evolution has transformed the flower petal into a precision optical and biomechanical instrument, finely tuned to the sensory capabilities of its primary pollinator, whose visual system is acutely sensitive to features like symmetry, contrast, and subtle reflection range.</p><p>(3) Floral Electric Fields: The Invisible Signal</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/746/1*gVuRBivxtYmmxLGQtftJUQ.png" /></figure><p>Floral electric fields represent an overlooked yet pivotal sensory signal in plant-pollinator interactions, emerging as a key component of the co-evolutionary dynamics between flowering plants and pollinators like bees. These invisible fields are laid on three core explanations: the atmospheric potential gradient (APG), the triboelectric charging of bees, and the inherent electrical properties of plants.</p><p>The APG, a vertical electric field spanning the Earth’s surface and the upper atmosphere (typically ~100 V/m in fair weather), provides the foundational environmental context. As grounded organisms, plants accumulate negative charge via electrostatic induction in response to the APG’s positive atmospheric potential.</p><p>This charge distribution is not uniform across floral structures — sharp features such as petal edges, stamens, stigmas, and trichomes exhibit higher charge density due to their geometry, especially charge density is inversely proportional to surface curvature.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/816/1*HLijxEUwrOzuIRfIy8ffNw.png" /></figure><p>Finite-element modeling and electrostatic dusting experiments have visualized these heterogeneous fields, revealing that floral anatomy shapes their spatial structure, much like how petal color or shape signals resource availability.</p><p>Bees, in contrast, acquire positive charge through the triboelectric effect — frictional contact with air, floral surfaces, or their own body parts during flight.</p><p>Experiments placing bees on the triboelectric series confirm they occupy the extreme positive end, with no tested material inducing a negative charge. This positive charge, ranging from +30 to +50 pC, interacts with the flower’s negative charge, generating intense electric fields as the bee approaches the flower. These fields are far stronger than the threshold required to trigger mechanical and neural responses in bees’ sensory structures.</p><p>For pollinators, floral electric fields function as a multimodal cue that complements vision, olfaction, and touch. Bumblebees detect these fields via mechanosensory hairs on their bodies, which deflect in response to electric forces — even at field strengths as low as 0.77 V/m.</p><p>Beyond sensory detection, floral electric fields play a direct role in pollen transfer. As a positively charged bee approaches a negatively charged flower, electrostatic forces between the two bodies intensify, overcoming gravity and viscous drag to guide pollen grains. Modeling shows that pollen moves bidirectionally: from the flower’s negatively charged anthers to the bee’s positive surface, and from the bee to the flower’s receptive stigma.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/350/1*XNEcXldrs28Fp4fzUHvKRA.png" /></figure><p>This graph shows how scientists used electrostatic dusting to make the invisible electric fields around flowers visible.</p><p>First, they took five different types of flowers — Lilium (lily), Gerbera, Narcissus (daffodil), Bergenia, and Petunia. For each flower, they took a photo <em>before</em> doing anything (the left side of each pair) and another <em>after</em> dusting it with colored powder that had an electrostatic charge (the right side; the powder was blue or yellow in the bottom image).</p><p>The way the powder stuck to the flowers wasn’t random. More powder piled up in some spots than others, and this tells us something about the flower’s electric charge: the areas with more powder are where the flower had a stronger negative charge.</p><p>What makes floral electric fields a signal rather than a passive physical phenomenon is their adaptiveness. Like colors, fragrances, or nectar rewards, these fields are shaped by evolutionary pressures. Environmental factors such as humidity or wind modulate the efficacy of these signals, adding complexity to their ecological function.</p><p><strong>III. What Is ‘Bee Purple’?</strong></p><p>‘Bee Purple’ is a term used to describe a specific color that is highly visible and attractive to bees, but which the human eye cannot fully perceive. It is not a single, distinct color like red or blue, but rather a perceptual blend that is unique to the bee’s visual system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/744/1*_GcXrBxQLkkCpypxZSX2Lw.png" /></figure><p>Bees are trichromatic, meaning they see color by combining input from three types of color-sensitive photoreceptors, just like humans. However, the wavelengths they see are shifted. Humans see combinations of red, green, and blue light. On the contrary, bees see combinations of Ultraviolet (UV), Blue, and Green light. They lack the receptor for red light, which appears dark or black to them.</p><p>Bee purple is a color perceived by the bee’s brain when its UV-sensitive receptor and its Blue-sensitive receptor are stimulated simultaneously.</p><p>Many flowers, such as Evening Primrose and Black-Eyed Susan, have petals that absorb visible light but strongly reflect UV light in specific patterns. To a human, the flower might look plain yellow, but to a bee, the UV-reflecting area combines with the visible blue or yellow light to create patterns of “bee purple.” These patterns often form a nectar guide, directing the bee straight to the pollen and nectar reward. Because flowers rich in nectar often display colors in the blue and purple range (which includes the UV-blue combinations), bees are genetically and behaviorally primed to be highly attracted to this color.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*jaYXGwG2gt-Otz13W2zOxA.png" /></figure><blockquote><strong>What we see is never all that exists.<br> The world is full of hidden colors, patterns, and signals, waiting for the right eyes to find them.</strong></blockquote><p>Reference:</p><p><a href="https://livebeekeeping.com/bees/how-bees-see/">https://livebeekeeping.com/bees/how-bees-see/</a></p><p><a href="https://www.cam.ac.uk/research/news/petals-produce-a-blue-halo-that-helps-bees-find-flowers">https://www.cam.ac.uk/research/news/petals-produce-a-blue-halo-that-helps-bees-find-flowers</a></p><p><a href="https://mrplantgeek.com/2022/07/16/why-do-bees-love-purple/#:~:text=July%2016%2C%202022,invisible%20to%20the%20human%20eye">https://mrplantgeek.com/2022/07/16/why-do-bees-love-purple/#:~:text=July%2016%2C%202022,invisible%20to%20the%20human%20eye</a>.</p><p><a href="https://beeculture.com/bees-see-matters/#:~:text=This%20is%20the%20reason%20why,of%20yellow%20and%20ultraviolet%20light">https://beeculture.com/bees-see-matters/#:~:text=This%20is%20the%20reason%20why,of%20yellow%20and%20ultraviolet%20light</a>.</p><p><a href="https://www.mannlakeltd.com/blog/what-does-a-bee-see/#:~:text=This%20characteristic%20is%20commonly%20shared,Like%20humans%2C%20bees%20are%20trichromatic">https://www.mannlakeltd.com/blog/what-does-a-bee-see/#:~:text=This%20characteristic%20is%20commonly%20shared,Like%20humans%2C%20bees%20are%20trichromatic</a>.</p><p><a href="https://marisamorby.com/bee-vision/#:~:text=This%20video%20is%20also%20really,them%20to%20come%20and%20land">https://marisamorby.com/bee-vision/#:~:text=This%20video%20is%20also%20really,them%20to%20come%20and%20land</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6bb878c262eb" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>